Computers Gone Wrong
I’m always happy to see evidence that the collective “we” are not as smart as is thought. The best and brightest climatologists (and their big computers) missed a big call on global weather patterns this fall.
This chart plots the expectations for the progress of El Nino as of July. The spaghetti strings are all over the lot, but the consensus was for a return to El Nino conditions sometime in October.
The evidence for this developing shift gave cause for NOAA to issue alerts starting in July.
The news media jumped on the forecast:
The El Nino alerts continued through October:
But it did not happen. Mother nature fooled the computers. There is no El Nino, and the odds for one developing have fallen substantially. The latest status report from NOAA, and more spaghetti strings:
(click to enlarge)
Does it matter if the scientist had it wrong? Not really. What has happened in the past month with weather would have happened regardless of the accuracy of the forecasts. But it would have (probably) made a very big difference if the forecasts had happened to be right.
When there are El Nino conditions, tropical storms tend have their “tops” sheared apart by winds that blow east. In addition, storms are pushed out into the Atlantic before they make landfall. This pic tells the story:
So if we had been in El Nino conditions on October 29, Sandy would never have gotten as big as it did, and might just have blown out to sea. What a difference that would have made.
I think that smart guys and gals should continue to use big computers to forecast the future. It’s helpful to look ahead with some rational expectations of what should happen next. That’s true for weather, stock prices and macro economic trends/performance.
But we should also look askance at what the computers and sages are telling us. The machines, and their operators, are consistently wrong. The bad news is that unanticipated events almost always have negative outcomes. The good news is that we can’t foretell the future; if we could, it wouldn’t be interesting at all.